These authors contributed equally to the work.
Genomic organization, alternative splicing, and multiple regulatory regions of the zebrafish fgf8 gene
Article first published online: 7 SEP 2006
Development, Growth & Differentiation
Volume 48, Issue 7, pages 447–462, September 2006
How to Cite
Inoue, F., Nagayoshi, S., Ota, S., Islam, Md. E., Tonou-Fujimori, N., Odaira, Y., Kawakami, K. and Yamasu, K. (2006), Genomic organization, alternative splicing, and multiple regulatory regions of the zebrafish fgf8 gene. Development, Growth & Differentiation, 48: 447–462. doi: 10.1111/j.1440-169X.2006.00882.x
- Issue published online: 7 SEP 2006
- Article first published online: 7 SEP 2006
- Received 22 May 2006; revised 27 June 2006; accepted 1 July 2006.
- conserved non-coding sequence;
- splicing variant;
- transcriptional regulation;
Fgf8 is among the members of the fibroblast growth factor (FGF) family that play pivotal roles in vertebrate development. In the present study, the genomic DNA of the zebrafish fgf8 gene was cloned to elucidate the regulatory mechanism behind the temporally and spatially restricted expression of the gene in vertebrate embryos. Structural analysis revealed that the exon–intron organization of fgf8 is highly conserved during vertebrate evolution, from teleosts to mammals. Close inspection of the genomic sequence and reverse transcription–polymerase chain reaction analysis revealed that zebrafish fgf8 encodes two splicing variants, corresponding to Fgf8a and Fgf8b, among the four to seven splicing variants known in mammals. Misexpression of the two variants in zebrafish embryos following mRNA injection showed that both variants have dorsalizing activities on zebrafish embryos, with Fgf8b being more potent. Reporter gene analysis of the transcriptional regulation of zebrafish fgf8 suggested that its complicated expression pattern, which is considered essential for its multiple roles in development, is mediated by combinations of different regulatory regions in the upstream and downstream regions of the gene. Furthermore, comparison of the genomic sequence of fgf8 among different vertebrate species suggests that this regulatory mechanism is conserved during vertebrate evolution.
Fibroblast growth factor (FGF) was originally identified as a potent mitogen for cultured fibroblasts (Gospodarowicz 1974). Since that discovery, a variety of roles have been revealed for FGF in embryonic development and in adult organisms of both vertebrates and invertebrates, and, to date, 22 FGF have been identified in mice and humans (Ornitz & Itoh 2001; Bottcher & Niehrs 2005). Fgf8 was originally identified as a growth factor that was responsible for the androgen-dependent growth of a mammary carcinoma cell line (Tanaka et al. 1992). It is now widely accepted that Fgf8 is among the most important growth factors regulating vertebrate embryogenesis.
The expression patterns of fgf8 are temporally and spatially highly restricted and dynamic in embryos, providing organizing signals for patterning of the embryonic body and various organs. From the blastula to the gastrula stage, fgf8 is expressed in precursors of endomesoderm, where it promotes mesoderm formation and the induction of neural tissues (Meyers et al. 1998; Sun et al. 1999; Mathieu et al. 2004; Rentzsch et al. 2004). From the gastrula to the somitogenesis stage, fgf8 is expressed in the anterior hindbrain of the neural plate, with a sharp anterior boundary that is considered to be the prospective midbrain–hindbrain boundary (MHB). The MHB region is regarded as the organizing center of brain formation, and Fgf8 produced therein is required for the establishment and maintenance of the MHB signaling center as well as for its organizing activity (Mason et al. 2000; Reifers et al. 1998; Rhinn & Brand 2001; Chi et al. 2003). In addition, Fgf8 is known to function synergistically with Shh to promote the formation of dopaminergic neurons in the ventral midbrain and serotonergic neurons in the ventral hindbrain (Ye et al. 1998; Teraoka et al. 2004). During somitogenesis and at later stages, fgf8 is expressed in the anterior marginal region of the prospective forebrain and is involved in the formation of the telencephalon and anterior diencephalon (Shimamura & Rubenstein 1997; Shinya et al. 2001; Kobayashi et al. 2002). In zebrafish embryos, the growth factor gene is expressed in the fourth rhombomere (r4), where it contributes to the patterning of the posterior hindbrain together with other fgf genes (Walshe et al. 2002). Furthermore, fgf8 is expressed in the posterior end of the neural tube and tailbud, thereby patterning the neural tube along the dorso-ventral and antero-posterior axes (Diez del Corral et al. 2003; Novitch et al. 2003; Dubrulle & Pourquie 2004). Recently, a new role has been proposed for fgf8; this gene is probably essential for maintenance of the spinal cord stem zone that progressively gives rise to the spinal cord and paraxial mesoderm (Delfino-Machin et al. 2005).
Outside the central nervous system, fgf8 is expressed in the unsegmented paraxial mesoderm, which determines the position where somitic vesicles are finally formed (Sawada et al. 2001), and then expressed in the anterior portion of forming somites, contributing to their anterior–posterior patterning (Durbin et al. 1998). This gene is also expressed in the otic vesicle and optic stalk, thereby regulating the development of these organs. In addition, fgf8 is expressed in the apical ectoderm ridge (AER) of the limb bud and patterns growth and differentiation of the limb bud by acting on the mesenchymal cells therein (Sun et al. 2002).
Since the functions and expression patterns of fgf8 are highly conserved among vertebrates, from teleost fish to mammals, elucidation of the regulatory mechanism of fgf8 expression in vertebrate embryos is essential for our understanding of vertebrate development and evolution. Transcriptional regulation has been analyzed using the gene-transfer technique in cultured embryonic stem (ES) cells and COS cells, and the results suggested the presence of an enhancer in one of the introns and another in the upstream DNA (Gemel et al. 1999; Brondani et al. 2002). However, due to the limitations inherent in this type of analyses, in which a simple cultured-cell system is used, little information has been obtained in terms of the spatial and temporal regulation of fgf8 within embryos during development. Only recently, Beermann et al. reported the presence of several sequences in the downstream region of mouse fgf8 that are conserved in other vertebrates, and showed that DNA fragments including some of these sequences partially recapitulate the fgf8 expression in mouse embryos (Beermann et al. 2006). However, the DNA fragments used were rather large as compared with the conserved sequences, and precise expression analysis of the transgene was not conducted, making it difficult to evaluate the suggested roles of the conserved regions.
Another important issue regarding the role of fgf8 in animal development is the physiological significance of the multiple splicing variants that have been found in a wide variety of vertebrates. To date, seven and four transcriptional splicing variants are known for mouse and human fgf8 (Crossley & Martin 1995; Gemel et al. 1996), respectively, while two variants have been identified in chicks and Xenopus (Sato et al. 2001; Shim et al. 2005). Several studies have shown that the transforming effects on NIH3T3 cells, tumorigenicity, and activities on respective FGF receptors differ among the seven isoforms of mouse Fgf8, although functional diversification is not well understood (see Discussion). In zebrafish, only cDNA clones for a single Fgf8 isoform have been reported, and the identification of other fgf8 isoforms is lacking.
In the present study, we show that the organization of this gene is highly conserved during vertebrate evolution, and that zebrafish fgf8 encodes two splicing variants, Fgf8a and Fgf8b, both of which have similar dorsalizing activities, but with different intensities, on zebrafish embryos. We further show that fgf8 expression in embryos is governed by combinations of different spatial cis-regulators that are distributed upstream and downstream to the gene.
Materials and methods
Adult zebrafish (Danio rerio) were maintained at 27°C in a 14:10 h light : dark cycle. Embryos were raised at 28.5°C until appropriate stages. Morphological features and hours postfertilization (h.p.f.) were used to stage embryos (Kimmel et al. 1995). When necessary, 0.2 mm phenylthiourea (Nakarai Tesque, Kyoto, Japan) was added to the cultures to prevent pigment formation.
Cloning of the genomic DNA for fgf8
A zebrafish genomic phage library (λFIXII, kindly donated by Dr H. Okamoto, BSI, Riken) was screened by plaque hybridization with the fgf8 full-length cDNA as a probe (Fürthauer et al. 1997). The upstream DNA of fgf8 was cloned by screening a zebrafish genomic BAC library (‘Down-to-the-Well’ zebrafish BAC DNA Pools; Incyte Genomics, Palo Alto, CA, USA) by polymerase chain reaction (PCR) using the two primers for the 3′ untranslated region: 5′-ATAGGTAAGCGAAGCACAGGATG-3′ and 5′-ACCACGGTAGGAAACCTGGGATA-3′. The BAC clone identified as positive (BAC168F14) was purchased from Incyte Genomics.
Reverse transcription–polymerase chain reaction
Total RNA was purified from embryos as described previously (Chomczynski & Sacchi 1987) and reverse transcribed by M-MLV reverse transcriptase (Gibco BRL, Gaithersburg, MD, USA). The resulting cDNA was subjected to PCR using the sense primer designed for exon 1a (Zfg8-s, 5′-CATGAGACTCATACCTTCAC-3′) and either of the two antisense primers for the 3′ portion of exon 1d (Zfg8-as2, 5′-GTTGTTTTCTTGTTGGCGAG-3′Fig. 1C) and exon 3 (Zfg8-as, 5′-TGTAGTTGTTCTCCAGGACT-3′). As an internal control, expression of the zebrafish gapdh gene (AY818346) was examined likewise using 5′-GGAATCAATGGATTTGGCCG-3′ and 5′-AGGCTTCATGCACTGGAACA-3′ as primers. In a separate PCR, the entire coding region of the fgf8 cDNA was amplified using sense and antisense primers (5′-GCGGATCCAAACATGAGACTCATACCTTCA-3′ and 5′-GCTCTAGATCAACGCTCTCCTGAGTAGC-3′, respectively; the sequence tags for the restriction enzymes are underlined). The resulting PCR products were cloned into pBluescript SK- and subjected to sequence analysis. LA-Taq polymerase (Takara, Otsu, Japan) was used in PCR (LA-PCR) throughout this study when sequences were critical.
Determination of the transcriptional initiation site
Total RNA was purified from 24-h.p.f. embryos as described above and used for 5′ rapid amplification of cDNA ends using the 5′ RACE system for Rapid Amplification of cDNA Ends (Gibco BRL, Gaithersburg, MD, USA) according to the manufacturer's protocol. In this process, cDNA was obtained from purified total RNA using GSP1 primer (5′-AAACTCGCGGCATGTAAACG-3′), and the resultant cDNA was amplified by nested PCR using two antisense gene-specific primers: GSP2, 5′-CAAGTTTAAGCGAAAGCGCC-3′, and GSP3, 5′-TAATCATGCAGCCCGAGGAA-3′.
Percent Identity Plot analysis
The genomic sequence of zebrafish fgf8 from −3768 to +17797 bp relative to the transcriptional start site was compared with corresponding regions of fgf8 from different vertebrates by Percent Identity Plot (PIP) analysis (Schwartz et al. 2000). In this sequence, the region from −3768 to −2501 (−3768/−2501) was obtained from Ensembl Genome Data Resources (Birney et al. 2006; scaffold 1152.2), the −2500/−289 and −288/+6567 regions were determined in BAC168F14 and λZfg8.2, respectively (deposited in the DDBJ/EMBL/GenBank databases with the accession number AB256522), and the sequence of the +6568/+17797 region was from the database (AB037997). The genomic sequences of Tetraodon nigroviridis (scaffold SCAF13770, −10.6 to +20.3 kb), Fugu rubripes (scaffold 778, −37.0 to +75.6 kb), Xenopus tropicalis (scaffold 708, −11.2 to +37.3 kb), Gallus gallus (contig 22.299, −20.0 to +36.3 kb), Mus musculus (Fgf8, contig AC149086.4, −30.2 to +25.9 kb), and Homo sapiens (FGF8, contig AC0107188.8.131.52842, −20.2 to +35.6 kb) were also obtained from Ensembl Genome Data Resources. The corresponding sequence of Oryzias latipes (scaffold 3522) was obtained from the Medaka Genome Project, provided by the NIG DNA Sequence Center (2004).
Preparation of mRNA for fgf8
The cDNA for Fgf8a and Fgf8b, which had been amplified by PCR in the present study, were cloned into the multicloning site (MCS) of pCS2+, and used as templates to synthesize mRNA, as described previously (Kikuta et al. 2003). The mRNA was solubilized in sterilized water and pressure-injected into one- to four-cell-stage embryos, which were allowed to further develop to appropriate stages.
The promoter region of fgf8 was excised from the genomic DNA in λZfg8.2 with SacI (–287) and AvaI (+181) and cloned into the SacI/SmaI site in the MCS of pEGFP-1 (Clontech, Palo Alto, CA, USA)(pZf8p-EGFP; Fig. 4B). During this procedure, the AvaI end was blunted before digestion with SacI. DNA containing the upstream region of 3.8 kb and the entire 5′ UTR (R[-3.8/5′UTR]) was amplified from pBAC-Nco by PCR using a T3 primer (5′-AATTAACCCTCACTAAAGGG-3′) for the MCS portion of pBAC-Nco and the downstream primer for the 3′ end of the 5′ UTR. After blunting with T4 DNA polymerase, the MCS sequence at the 5′ end of the PCR product was removed with EcoRI and cloned into the EcoRI/SmaI site of pEGFP-1 (pFg8(-3.8)-EGFP). The S2 region (+5.1 to +10.6 kb) was excised from the genomic DNA with SacI and inserted into the SacI site in pZf8p-EGFP in either the forward or reverse orientation (pS2(+)-EGFP or pS2(–)-EGFP, respectively). The DCR1/2 and S4.2 regions (Fig. 4A) were amplified from the genomic clone by LA-PCR using two pairs of primers; 5′-GGTACCACCTATTGACTCTC-3′/5′-GGCTTTCCAGCTGAATTGTC-3′ and 5′-AACTGGCATTCACTTCATTT-3′/5′-TGTGGCTAATCAAATGCTTCC-3′, and the DNA fragments obtained were ligated into the MCS of pEGFP-1 (pDCR1/2-EGFP and pS4.2-EGFP, respectively).
Introduction of reporter genes into embryos
DNA fragments in the GFP constructs harboring genomic DNA ligated to promoters and the egfp gene were PCR-amplified using a pair of primers designed for the two sites in pEGFP-1, one immediately upstream to the MCS (EGFP-5, 5′-AGTTATTACTAGCGCTACCGG-3′) and another downstream to the poly(A) addition site (EGFP-3, 5′-CGCCTTAAGATACATTGATGA-3′). The amplified DNA was injected into one-cell stage embryos at 10–20 pg/embryo.
Three types of promoter-EGFP DNA were used to examine the regulatory activities of genomic DNA fragments by co-injection: (i) Zf8p-egfp DNA, containing the fgf8 promoter region and the downstream egfp gene, was obtained by PCR amplification from pZf8p-EGFP using the primers EGFP-5 and EGFP-3 (Fig. 4B); (ii) the hsp-EGFP DNA fragment, which was excised from pzfHSP70/4-EGFP-pA (kindly donated by Dr H. Sasaki and Dr J. Kuwada), and consisted of the 1.5 kb promoter of zebrafish hsp70 (Halloran et al. 2000) and the downstream egfp; and (iii) ef1α-EGFP DNA, which comprised the promoter of the zebrafish ef1α gene and egfp, and was excised from the plasmid pzfEF1α-EGFP-pA (donated by Dr H. Sasaki and Dr M. Furutani-Seiki). In co-injection experiments, promoter-EGFP DNA was injected at 10 pg/embryo into one-cell embryos together with a three- to fivefold molar excess of the genomic fragments of interest.
DNA fragments for injection were fractionated by agarose-gel electrophoresis and extracted from the gel using the Qiaex II Gel Extraction Kit (Qiagen, Hilden, Germany). The DNA was solubilized in sterilized water and pressure-injected into embryos, which were allowed to develop to appropriate stages. GFP expression was determined by assessing the resultant fluorescence under a fluorescence stereomicroscope equipped with a GFP2 filter (MX FLIII; Leica, Wetzlar, Germany)
Establishment of transgenic lines
The embryos that were injected with plasmid constructs and expressed GFP fluorescence around 24 h.p.f. were allowed to develop further to maturation. Following mating with each other or with wild-type fish, as assessed by GFP fluorescence in the offspring embryos, founder transgenic fish (Tg fish, F0) that harbored the constructs in their germ lines were identified. The offspring (F1) obtained, which showed GFP fluorescence, were selected to generate Tg fish lines.
Whole-mount in situ hybridization
Digoxigenin (DIG)-labeled RNA probes were synthesized with T3 or T7 RNA polymerases (Stratagene, La Jolla, CA, USA) using the DIG RNA Labeling Mix (Roche Diagnostic, Indianapolis, IN, USA) according to the manufacturers’ protocols. Whole-mount in situ hybridization was performed essentially as described previously (Schulte-Merker et al. 1992; Kikuta et al. 2003).
Genomic organization and splicing variants of the zebrafish fgf8 gene
Two lambda phage clones and a BAC clone of the zebrafish fgf8 gene were obtained. Comparison of the genomic sequence with that of zebrafish fgf8 cDNA (Fürthauer et al. 1997; Reifers et al. 1998), which seems to encode the Fgf8b equivalent (see below), led to identification of exons 1a, 1b, 1d, 2, and 3, as is known in Fgf8 of mouse and human (Fig. 1A–C; Crossley & Martin 1995; Tanaka et al. 1995). The transcriptional start site (+1) was determined by 5′ RACE, and the transcribed region was found to span 6220 bp of genomic DNA. The larger phage clone (λZfg8.2) and a subclone from the BAC clone (pBAC-Nco) together covered the genomic region of fgf8 from −3.8 kb to +17.8 kb (Fig. 1A). At the 3′ end of this region is located the 5′-terminal coding sequence of hagoromo, which was identified as the causative gene of a stripe-pattern mutation encoding the zebrafish orthologue of Dactylin (Kawakami et al. 2000).
Four to seven splicing variants are known for fgf8 in mammals (Fgf8a-g; Crossley & Martin 1995; Tanaka et al. 1995; Gemel et al. 1996), two of which have also been identified in chicks and Xenopus (Fgf8a, b; Sato et al. 2001; Shim et al. 2005). In mammalian fgf8, alternative usage of exons 1b–d accounts for these complicated splicing patterns. Interestingly, the 5′-terminal 33 bp sequence of exon 1d (1d-5′) contains an AG sequence at its 3′ end that may function as the splicing acceptor (Fig. 1C), suggesting that zebrafish fgf8 encodes Fgf8a in addition to Fgf8b, as in other vertebrates. The presence of splicing variants for fgf8 in zebrafish embryos was examined by PCR of cDNA from bud-stage or 24 h.p.f. embryos between two sites within exon 1a and exon 3 whose amplification would yield all possible variants known to date (primers; Zfg8-s and Zfg8-as). Of the 10 partial cDNA clones obtained (five clones from the bud-stage and five from 24 h.p.f. embryos), nine represented the previously reported zebrafish Fgf8, while another, from the 24 h.p.f. embryos, encoded an Fgf8 isoform that lacked the 11 amino acids encoded by the 5′-terminal 33 bp of exon 1d. This 11 amino acid (11-aa) sequence is highly similar to the corresponding regions of mouse and human Fgf8b (82% to mouse and human Fgf8b). Based on the splicing pattern and sequence similarity, it was concluded that the newly identified and previously reported Fgf8 isoforms correspond to zebrafish Fgf8a and Fgf8b, respectively (Fig. 1C). PCR was also carried out to obtain the full-length fgf8 cDNA. Fourteen clones were obtained, among which 12 and two clones encoding Fgf8b and Fgf8a, respectively. To determine the temporal regulation of the two Fgf8 isoforms during zebrafish development, the levels of mRNA for the two isoforms were examined by reverse transcription–polymerase chain reaction (RT–PCR). No maternal expression was detected for either isoform, while zygotic mRNA appeared at the sphere stage and increased in amount later in development. Consistent with the ratio of the PCR-product clones obtained, Fgf8b was significantly predominant, and the ratio of the expression levels of the two isoforms was almost constant (Fig. 2A). Taken together, our results indicate that zebrafish fgf8 encodes at least two splicing variants, Fgf8a and Fgf8b, as do the chick and Xenopus fgf8 genes, and that Fgf8b is the dominant isoform during zebrafish embryogenesis.
The activities of the different Fgf isoforms have been compared in different assay systems, suggesting differences in their activities in various biological aspects (see Discussion). In the present study, Fgf8a and Fgf8b mRNA were introduced into zebrafish embryos, and dorsalization of embryos at the bud stage, which is a common activity of fgf genes in zebrafish embryos (Fürthauer et al. 1997; Reifers et al. 1998; Fürthauer et al. 2001), was examined. Both Fgf8 variants caused elongation of embryos (Fig. 2B, Table 1), a typical feature of dorsalization in zebrafish embryos (Fürthauer et al. 1997). In fact, expression of dorsal marker genes specific to the forebrain (six3b; Kobayashi et al. 1998; Nornes et al. 1998), MHB (pax2a; Krauss et al. 1991), and hindbrain (egr2b/krox20; Oxtoby & Jowett 1993) expanded ventrally, confirming that the Fgf8a identified here is in fact functional, and that zebrafish Fgf8a has a dorsalizing activity as do Fgf8b and other Fgfs. However, the effective dose was much higher (Table 1) while the extent of ventral expansion of dorsal marker expression was less significant for Fgf8a than for Fgf8b (Fig. 2C), showing that zebrafish Fgf8b is more potent at least in terms of dorsalization of embryos.
|Gene||Isoform||mRNA (pg/embryo)||n‡||Dorsalized (%)§|
Comparison of the fgf8 genomic sequences among vertebrate species
Numerous genomic sequences from a variety of animal species are now available. Comparative genomic studies employing these accumulated data have revealed that vertebrate genomes contain a large number of highly conserved sequences in their non-coding regions, and that some of these sequences act as transcriptional regulatory elements (Sandelin et al. 2004; Siepel et al. 2005; Woolfe et al. 2005). In the present study, the 22 kb flanking sequence of zebrafish fgf8 (−3.8 to +17.8) was compared with those of other vertebrates, including three other teleost fishes, frog, chick, and mammals, by the PIP analysis (Fig. 3). Five of the known fgf8 exons (exons 1a, 1b, 1d, 2, and 3) were located in the fgf8 genes of all the species by sequence similarity and/or comparison with their cDNA sequences. However, since the fgf8 variants encoded by exon 1c are known only in mice and humans (Crossley & Martin 1995; Gemel et al. 1996), it was impossible to establish whether exon 1c is also present in the fgf8 genes of vertebrates other than mammals.
In the upstream DNA, one sequence was found to be conserved in all the teleost fgf8 genes examined (upstream conserved region 1, UCR1, from −465 to −399; positions referring to those in zebrafish fgf8 hereafter), but was not conserved in amniotes and frogs. The sequence in the vicinity of the promoter region (conserved promoter region; CPR) is also conserved only in teleosts, while highly GC-rich regions are present immediately upstream to the fgf8 genes only in amniotes (Fig. 3). It was reported previously that an intronic region located between exon 1d and exon 2 is conserved between the mouse and human, and that this region acts as an enhancer in ES cells (Gemel et al. 1999). We detected this intronic conserved region also in the chick and Xenopus. Furthermore, its central portion with the size of 165 bp is present even in three of the fish species examined, including zebrafish (intronic conserved region; ICR). At the 3′-end of the transcribed region, a 107 bp sequence (terminal conserved region; TCR) was found in all the teleosts examined as well as in Xenopus.
In the downstream region, a 258 bp sequence is conserved among zebrafish, Tetraodon, and Fugu, but not in medaka and Xenopus (downstream conserved region 1, DCR1, at +10.3 kb). The 5′-portion of DCR1 (54 bp, DCR1a) is also present in amniotes at equivalent positions, while additional flanking sequences are conserved only in amniotes (DCR1a-5 and DCR1a-3). The 340 bp DCR2 region, at +10.8 kb of zebrafish fgf8, is conserved at equivalent sites in Tetraodon and Fugu, in which small gaps separate DCR2 into two subregions (DCR2a and DCR2b), but is absent in other vertebrate species, including medaka. Unexpectedly, the third conserved region, at +14.6 kb of zebrafish fgf8 (DCR3, 309 bp), was not found in other fish or in Xenopus, but was conserved in chicks and mammals, in which additional flanking regions of DCR3 were also found to be conserved (DCR3-5 and DCR3-3). Some of these findings, especially the conservation of DCR3 only in zebrafish and amniotes, are not consistent with predictions based on the phylogenetic relationship, an issue that is discussed below. We also identified four non-coding conserved regions (additional conserved regions; ACR1-4), which are conserved in some vertebrate species, but not in zebrafish.
Multiple regulatory sequences in the flanking region of fgf8
In order to understand the complicated and highly organized regulation of fgf8, we examined the transcriptional regulatory activities of the flanking DNA. The region from −3.8 kb to +17.8 kb was divided into several subregions (R[-3.8/5′UTR], X1/X1*, XS1, S2, S3, S4, S5; Fig. 4A), and their activities, as measured by the expression of the GFP reporter gene, were assayed in injected embryos. Among these subregions, R[-3.8/5′UTR] (−3.8 to +0.5 kb), X1/X1* (−0.3 to +2.1/2.2 kb), S2 (+5.1 to +10.6 kb), S3 (+10.6 to +13.6 kb), and S4 (+14.0 to +17.8 kb) contain UCR1, ICR, TCR/DCR1, DCR2, and DCR3, respectively. Since transient expression tends to be mosaic and to generate ectopic expression, only robust expression in a relatively broad area was scored.
The upstream R[-3.8/5′UTR] region, which includes the fgf8 promoter, was directly ligated to egfp (Fg8(-3.8)-EGFP). The results showed that this region is able to drive expression in the surface ectoderm, especially the facial ectoderm and prospective caudal fin region, at 24 h.p.f. (Fig. 5A, Table 2). Surface expression similar to that of Fg8(-3.8)-EGFP, shown here, was described previously for zebrafish fgf8 (Reifers et al. 1998; Yonei-Tamura et al. 1999; Zebrafish Information Network). We then co-injected into one-cell embryos each of the other subregions with Zf8p-EGFP, in which the egfp gene is regulated by the fgf8 promoter (−288 to +180, Fig. 4B). As shown in Table 3, green fluorescent protein (GFP) fluorescence was thus observed in developing embryos. It is known that co-injected regulatory DNA can drive reporter genes, probably due to rapid in ovo ligation after injection (Muller et al. 1997; Udvadia & Linney 2003). This strategy has been employed successfully in zebrafish to demonstrate the regulatory functions of several non-coding genomic regions highly conserved in vertebrates (Woolfe et al. 2005). Preliminary experiments confirmed that Zf8p-EGFP alone caused little GFP expression, despite the presence of the conserved promoter region (CPR) around the initiator site (Figs 3 and 4, Table 2).
|Zf8p-EGFP||48||2 (4)||1 (2)||0 (0)||0 (0)||1 (2)|
|Fg8(-3.8)-EGFP||55||54 (98)||0 (0)||0 (0)||0 (0)||45 (82)|
|S2(+)-EGFP||67||65 (96)||27 (40)||2 (3)||5 (7)||30 (45)|
|S2(–)-EGFP||103||96 (93)||67 (65)||8 (8)||8 (8)||41 (40)|
|(A) 8–10 somite stage|
|Genomic region||n‡||Expression (%)§|
|–¶||191||6 (3)||0 (0)|
|X1*‡‡||26||22 (85)||0 (0)|
|XS1||43||42 (98)||0 (0)|
|S2||104||83 (80)||2 (2)|
|S3||37||35 (95)||16 (43)|
|S5||37||24 (65)||0 (0)|
|S4||66||66 (100)||9 (14)|
|(B) 24 h.p.f.|
|Genomic region||n‡||Expression (%)§|
|–¶||48||2 (4)||1 (2)||0 (0)||0 (0)||1 (2)|
|X1‡‡||54||33 (61)||2 (4)||2 (4)||0 (0)||13 (24)|
|XS1||53||52 (98)||13 (25)||10 (19)||1 (2)||20 (38)|
|S2||64||63 (98)||28 (44)||4 (6)||9 (14)||37 (58)|
|S3||22||16 (73)||3 (14)||2 (9)||1 (5)||10 (45)|
|S5||42||3 (7)||0 (0)||0 (0)||1 (2)||1 (2)|
|S4||43||42 (98)||7 (16)||6 (14)||14 (33)||23 (53)|
At early somitogenesis, only S3 DNA yielded significantly high expression, which occurred in a broad region of the brain, including the midbrain and hindbrain (Fig. 5C, Table 3A). In addition, the regulatory function of this region was confined to the 687 bp at the 5′-end of S3, from +10.6 to +11.3 kb (S3.1; Fig. 4A, and data not shown), which includes the DCR2 region. Around the end of somitogenesis, the S2 region (Fig. 4A) gave rise to high expression in the anterior head ectoderm, especially in the posterior–ventral telencephalon (Fig. 5B, Table 3B). Meanwhile, the S4 region (Fig. 4A) directed GFP expression in the MHB, optic stalk, and otic vesicles (Fig. 5D, Table 3B, and data not shown), all of which are known to express endogenous fgf8 in zebrafish embryos (Fürthauer et al. 1997; Reifers et al. 1998). The XS1 region (+2.1 to +5.1 kb; Fig. 4A) also drove spatially restricted expression to a certain extent (Table 3B), although the expression domain was less distinct (data not shown), and thus this region was not examined further in the present study. Essentially the same expression patterns were obtained when each of the subregions was co-injected with egfp that was under regulation by the heat shock promoter hsp70/4 (Halloran et al. 2000) or the ef1a promoter, or when injected as GFP constructs in which the corresponding genomic regions were ligated to Zf8p-EGFP (Table 2; and data not shown). In the context of the plasmid constructs, the S2 region drove the reporter gene in a spatially specific manner in either orientation relative to egfp (Table 2; S2(+/–)-EGFP).
Transient expression assays offer rapid analysis of functional cis elements scattered throughout large genomic DNA; however, the resulting mosaicism limits the use of this approach in precise determination of DNA regulatory functions. In order to circumvent this difficulty, we established transgenic (Tg) fish that passed the transgene through the germ line. Accordingly, seven lines of Tg fish were obtained for the S2-EGFP construct, all but one of which generated essentially the same expression pattern. The results of experiments in the Tg fish again demonstrated that the regulatory function of S2 is independent of its orientation (data not shown). S2-EGFP Tg embryos showed distinct GFP expression in the posterior telencephalon, consistent with the results of the transient assay (Fig. 5E,F). The embryos also specifically expressed GFP in the caudal spinal cord, tail bud, heart, and optic stalk. All of these embryonic regions except for the posterior telencephalon coincide with or include as parts the sites that are known to express fgf8 in many vertebrates, including zebrafish (Crossley & Martin 1995; Reifers et al. 2000; Dubrulle & Pourquie 2004; see also Fig. 6C,D,F,H,K,L).
It should be noted that, in contrast to fgf8 mRNA expression in zebrafish, the GFP domain in the spinal cord was much larger in the Tg fish at 24 h.p.f. In the former, expression was confined to the caudal end of the spinal cord (data not shown; see also Fig. 6C,D,K). Nonetheless, egfp transgene mRNA expression was also restricted to a similar caudal region in S2-EGFP Tg embryos (Fig. 6A,B,I), suggesting that broad expression of GFP in the spinal cord is due to the well-known stable nature of GFP (Li et al. 1998). Detection of the egfp mRNA confirmed the expression of the transgene in the posterior telencephalon, optic stalk, heart, and tail bud. Furthermore, in addition to the striking expression of fgf8 in the anterior dorsal telencephalon (asterisks) and dorsal diencephalon (black triangles) (Fig. 6F,H,L; Fürthauer et al. 1997; Reifers et al. 1998), we detected weak, but distinct expression of endogenous fgf8 in the posterior telencephalon in wild type embryos (Fig. 6F,H,L, open triangles), to which S2 directed reporter expression (Fig. 6E,G,J, open triangles). Thus, it appears that S2 recapitulates the endogenous expression of fgf8 in the caudal spinal cord, tail bud, optic stalk, posterior telencephalon, and heart. Further analysis of the S2 region showed that the 3′-region of S2 (S2.2, +10.1 to +10.6 kb; Fig. 4A), which contains the DCR1 region, is responsible for expression in the caudal spinal cord and posterior telencephalon (Odaira, Inoue and Yamasu; unpubl. data).
We also established two Tg lines that harbored GFP constructs (DCR1/2-EGFP), in which egfp was driven by the DCR1/2 region (from +10.1 to +11.1 kb) including both DCR1 and DCR2. In embryos of both these lines, DCR1/2 drove GFP expression in a broad region of the brain region at early somitogenesis, from the midbrain to the hindbrain (Figs 5G,6M), in addition to expression in caudal neural tube and telencephalon, which was attributable to the activity of the S2.2 region. The broad expression of GFP in the midbrain and hindbrain in Tg embryos was quite similar to the transient GFP expression due to S3/S3.1 (Fig. 5C), showing that the S3.1 region, which contains the DCR2 region, drives broad expression in the embryonic brain. In zebrafish, fgf8 is expressed in the anterior hindbrain at the late gastrula stage, and this expression gradually splits to three anterior rhombomeres, r1, r2, and r4, at early somitogenesis (Reifers et al. 1998). Thus, S3-directed hindbrain expression of GFP seems to include these hindbrain subregions, while midbrain expression is apparently ectopic.
As to the S4 region, DNA fragments of the five subregions were prepared (Fig. 4A), which were examined for their regulatory activities, revealing that the regulatory activity of S4 is localized in the S4.2 region (+14.4 to +15.7 kb). We then established three lines of Tg fish harboring the GFP constructs (S4.2-EGFP), in which S4.2 was ligated to zf8p-EGFP. In all the lines obtained, the transgene was expressed during the somitogenesis stage in the MHB region, optic stalk, and otic vesicle (Figs 5H,6N, and data not shown), as was observed in the transient expression.
Taken together, the results of transgenic analysis supported and substantiated those from the transient assay, and further defined the precise patterns of transcriptional regulation.
Organization and alternative splicing of the zebrafish fgf8 gene
The organization of fgf8 in zebrafish, as presented here, shows that the basic structure of the gene is highly conserved among vertebrates, from fish to mammals. We further showed that zebrafish fgf8 encodes two splicing variants, Fgf8a and Fgf8b, as is known in other vertebrates (Crossley & Martin 1995; Sato et al. 2001; Shim et al. 2005). In the case of zebrafish fgf8, there is no evidence, either by genomic sequencing or RT–PCR analysis, of the presence of other variants, such as Fgf8c-h. This is consistent with the situation in chicks and Xenopus, in which only Fgf8a and Fgf8b have been identified (Sato et al. 2001; Shim et al. 2005). Although the possibility of additional Fgf8 isoforms cannot be excluded, we presently favor the view that the appearance of exon 1c, accompanied by the generation of more than two variants, occurred after the diversification of mammals from other tetrapod vertebrates, leading to refinement of the regulation of organogenesis by Fgf8 in the mammalian lineage. Thus, the significance of the increase in splicing complexity deserves further attention with regard to vertebrate evolution.
Several studies have reported differences in the biological activities of Fgf8 variants in a number of assay systems. The effects of Fgf8b in transforming NIH3T3 cells and promoting their tumorigenicity were stronger than those of Fgf8a and Fgf8c (MacArthur et al. 1995a). In misexpression experiments involving in ovo electroporation into chick embryos, Fgf8b was 100-fold more active than Fgf8a in transforming presumptive diencephalon to the tectum (Sato et al. 2001). In cells transfected with genes of different FGF receptor isoforms (FGFR), differences in the activation of respective FGFR were found among different Fgf8 variants (MacArthur et al. 1995b; Blunt et al. 1997). It was shown that FGFR3c (‘c’ splice form of FGFR3) and FGFR4 are activated by Fgf8b-g, while Fgf8b and 8f also efficiently activate FGFR2c; in contrast, Fgf8a has little effect on these receptors. These results were confirmed by recent studies using surface plasmon resonance technique (SPR) (Olsen et al. 2006).
The physiological significance of the multiple splicing variants remains enigmatic, and no clear differences have been observed in either the spatial or temporal regulation of the expression of seven Fgf8 isoforms during mouse development (Crossley & Martin 1995; MacArthur et al. 1995b; Blunt et al. 1997). However, possible different functions were recently observed for Fgf8a and Fgf8b regarding brain formation in chick embryos. Based on the results obtained in the misexpression experiments described above, it seems that Fgf8a induces formation of the midbrain anterior to the MHB, while Fgf8b promotes cerebellum formation in the posterior region (Sato et al. 2001). Also, in the mouse, misexpression of Fgf8a at MHB affected the growth and polarity of the forming midbrain (Lee et al. 1997), while Fgf8b converted the rostral brain to the hindbrain fate (Liu et al. 1999). These apparent functional differences are striking, although it is not clear whether they are attributable to differences in signaling intensities or pathways elicited by the different Fgf8 isoforms.
The results of the present study also showed that Fgf8b is the major Fgf8 isoform in zebrafish embryos throughout development, which is in keeping with the situation in other vertebrates (Crossley & Martin 1995; Sato et al. 2001; Shim et al. 2005). While dorsalizing activities were observed for zebrafish Fgf8a, as is known for zebrafish Fgf8b (Fürthauer et al. 1997), the activity of Fgf8a was significantly weaker than that of Fgf8b, which is consistent with the activities of Fgf8 isoforms in mouse and chick. It is therefore likely that the roles of the two isoforms in zebrafish development are compatible with those in amniote embryos. Recent crystallographic studies and SPR analysis showed that the functional difference between mouse Fgf8a and Fgf8b is attributable to a Phe residue in the 11-aa sequence, which is also present in zebrafish Fgf8b but absent in Fgf8a (Olsen et al. 2006). Interestingly, this residue is included in the N-linked glycosylation site that is also a common characteristic of Fgf8b from different vertebrate species (Fig. 1C; Crossley & Martin 1995; Shim et al. 2005). However, functional differences between Fgf8a and Fgf8b remain to be examined in zebrafish embryos.
Conserved sequences in the flanking region of fgf8
Recent studies based on accumulated genomic data from different species have established the presence of highly conserved sequences across the genome, many of which drive genes involved in developmental regulation (Sandelin et al. 2004; Siepel et al. 2005; Woolfe et al. 2005). Using the PIP analysis, we showed that multiple conserved non-coding sequences exist around fgf8. Moreover, as discussed below, our analysis of the regulatory functions of genomic DNA suggests that these conserved sequences indeed regulate fgf8 in zebrafish embryos. Some of the sequences identified here are conserved widely among vertebrate species (ICR, DCR1, DCR3), providing further support for their common role in the regulation of fgf8 expression required during vertebrate development. For this reason, the absence of these sequences in some lower vertebrates is puzzling considering their phylogenetic position. This is especially true for DCR3, which is present in zebrafish and amniotes, but not in other lower vertebrates. However, it should be taken into consideration that only a limited range of flanking sequences of respective species was examined, and many unknown sequences remain within the sequences of the lower vertebrates examined here. The missing sequences may be present in these as yet undetermined regions or they may have been translocated in the genome from fgf8 to much more distant sites during evolution.
Combinatorial functions of multiple regulatory sequences can give rise to the characteristic expression of fgf8 in embryos
Fgf8 is one of the most frequently described FGF in terms of regulating embryogenesis in vertebrates (Ornitz & Itoh 2001; Dono 2003; Bottcher & Niehrs 2005). It is expressed in several distinct embryonic regions during the course of development, thereby providing important signals that guide patterning of the entire body, brain, and other organs in embryos (see Introduction). The present reporter analysis identified regulatory genomic regions for expression during somitogenesis in MHB/r1, otic vesicle (S4), posterior telencephalon, heart, caudal spinal cord (S2), optic stalk (S2 and S4), and epidermis (upstream R[-3.8/5′UTR] region) (Fig. 7). A transient assay showed that the expression both in the posterior telencephalon and posterior spinal cord is driven by the 3′-most 0.5 kb region of S2 (S2.2; Fig. 4A), while the 5′-terminal 2.5 kb region of S2 activates transcription in the heart (Odaira, Inoue and Yamasu; unpubl. data). Since all of these expression domains are sites of fgf8 expression in zebrafish embryos, we propose that these flanking regions are indeed involved in the transcriptional regulation of fgf8. Broad GFP expression in the brain due to S3.1 was observed both in transgenic lines and in a significantly high proportion of embryos in the transient assays, making it likely that DCR2 directs expression in this brain region. Since this broad expression in the brain includes the anterior rhombomeres (r1, r2, and r4) where fgf8 is expressed at equivalent stages, there could be an additional tissue-specific silencer element(s) somewhere outside S3.1 that suppresses its ectopic enhancer activity in the midbrain, r3, and posterior rhombomeres. Finally, we have shown in both the transient and transgenic analyses that the S4.2 region drives expression in the MHB, optic stalk, and otic vesicle, recapitulating fgf8 expression in these regions during somitogenesis. Thus, the S4.2 region appears to be essential for the formation of the isthmic organizer/MHB that emanates Fgf8 and patterns the adjacent brain (Rhinn & Brand 2001). However, this region does not drive fgf8 at the late gastrula stage in the anterior hindbrain, where fgf8 expression contributes to the establishment of the MHB (Rhinn & Brand 2001). Detailed analysis of the S4.2 region is now in progress.
As discussed above, it is becoming widely accepted that conserved non-coding sequences act as functionally important regulatory regions. This notion has led to insights into the transcriptional regulation of many genes based on the genomic sequences. It should be noted that each of the three regions shown here to drive specific expression in zebrafish embryos (S2.2, S3.1, S4.2) contains a sequence that is conserved in other teleosts and/or even in higher vertebrates (DCR1, DCR2, and DCR3, respectively). The physiological relevance of these conserved regions remains to be revealed. In contrast, the role of the intronic conserved sequence, whose enhancer activity was demonstrated in a cultured cell system, was not apparent in our study. In fact, we did not observe strikingly specific function of the X1 and X1* regions, which include, respectively, the ICR partially and entirely. Still, these fragments drove expression in embryos to a certain degree, and the role of this region in embryos should be addressed in future studies by passing the transgene through the germ line. The XS1 region, which also partially includes the ICR, directed restricted expression in embryos, and its regulatory function is now under further analysis. The CPR is another region that is widely conserved in teleosts, although we were unable to identify its regulatory function. Furthermore, since the three promoters used in this study (fgf8, hsp70, and ef1a) gave consistent results in the transient assay (data not shown), the biological significance of the CPR is elusive at present. Beermann et al. recently reported the presence of several sequences in the downstream region of mouse fgf8 that are conserved in other vertebrates. Though the precise positions or sequences were not presented, the two conserved regions identified in their study, CR2 and CR3, may correspond to DCR1a and DCR3 in our study, respectively, based on the positions in the genome (Beermann et al. 2006). These investigators showed that the 2.5 kb mouse genomic region containing CR2 drives expression in tailbuds and isthmus, and that the 0.8 kb region containing CR3 regulates transcription in the AER, mandibular and maxillary arches, somites, and isthmus. While those results and our data are partially consistent, further analysis of the flanking genomic regions and the roles of respective conserved sequences in the mouse and zebrafish should be addressed in greater detail.
In the present study, the genomic DNA examined here did not recapitulate robust fgf8 expression either in the anterior hindbrain at the late gastrula or in the ANB/dorsal telencephalon and forming somites during somitogenesis. Thus, additional regulatory sequences may be located farther upstream or downstream to the genomic region examined here that drive fgf8 expression in these important signaling centers. It is also possible that reporter-gene expression in more restricted embryonic regions, where fgf8 is known to be expressed, escaped our observation in the transient analysis, which was used to efficiently, but roughly, survey a large extent of the flanking region. Detailed analyses in transgenic fish will be required to address this issue. Another problem to be addressed is why the downstream regulatory elements do not regulate the hagoromo gene, which is expressed ubiquitously in embryos (Kawakami et al. 2000) in spite of its close position to fgf8 in the genome.
Whatever the entire regulatory mechanism of fgf8, the present study has made it clear that the complicated pattern of fgf8 expression in a number of embryonic regions is achieved by a combination of multiple regulatory sequences distributed upstream and downstream to fgf8 (Fig. 7). This gene represents one of the typical synexpression groups, whose members are expressed in a highly similar and characteristic manner (Niehrs & Meinhardt 2002), and the regulatory mechanism shown here for fgf8 could be the prototype for this gene group.
We thank Drs M. Fürthauer, T. Johansen, T. Jowett, H. Sasaki, S. Schulte-Merker, H. Takeda, and D. Turner for providing the plasmids, and Dr H. Okamoto for providing the genomic library of zebrafish. We are also grateful to Dr M. Furutani-Seiki and Dr J. Y. Kuwada for allowing us to use the minimal promoters of zebrafish ef1a and zebrafish hsp70, respectively. This work was supported in part by the Ministry of Education, Culture, Sports, Science and Technology (Japan), Grants-in-Aid to K. Y. (nos 10220203, 15570170, 17570170).
- 2006. Identification of evolutionarily conserved regulatory elements in the mouse Fgf8 locus. Genesis 44, 1–6. , , , , &
- 2006. Ensembl 2006. Nucleic Acids Res. 34, D556–D561. , , et al.
- 1997. Overlapping expression and redundant activation of mesenchymal fibroblast growth factor (FGF) receptors by alternatively spliced FGF-8 ligands. J. Biol. Chem. 272, 3733–3738. , , , , &
- 2005. Fibroblast growth factor signaling during early vertebrate development. Endocr. Rev. 26, 63–77. &
- 2002. Promoter of FGF8 reveals a unique regulation by unliganded RARa. J. Mol. Biol. 319, 715–728. , , &
- 2003. The isthmic organizer signal FGF8 is required for cell survival in the prospective midbrain and cerebellum. Development 130, 2633–2644. , , &
- 1987. Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162, 156–159. &
- 1995. The mouse Fgf8 gene encodes a family of polypeptides and is expressed in regions that direct outgrowth and patterning in the developing embryo. Development 121, 439–451. &
- 2005. Specification and maintenance of the spinal cord stem zone. Development 132, 4273–4283. , , , &
- 2003. Opposing FGF and retinoid pathways control ventral neural pattern, neuronal differentiation, and segmentation during body axis extension. Neuron 40, 65–79. , , , , &
- 2003. Fibroblast growth factors as regulators of central nervous system development and function. Am. J. Physiol. Regul. Integr. Comp. Physiol. 284, R867–R881.
- 2004. fgf8 mRNA decay establishes a gradient that couples axial elongation to patterning in the vertebrate embryo. Nature 427, 419–422. &
- 1998. Eph signaling is required for segmentation and differentiation of the somites. Genes Dev. 12, 3096–3109. , , et al.
- 2001. sprouty4 acts in vivo as a feedback-induced antagonist of FGF signaling in zebrafish. Development 128, 2175–2186. , , , &
- 1997. A role for FGF-8 in the dorsoventral patterning of the zebrafish gastrula. Development 124, 4253–4264. , &
- 1996. Structure and sequence of human FGF8. Genomics 35, 253–257. , , &
- 1999. Fibroblast growth factor-8 expression is regulated by intronic engrailed and Pbx1-binding sites. J. Biol. Chem. 274, 6020–6026. , &
- 1974. Localisation of a fibroblast growth factor and its effect alone and with hydrocortisone on 3T3 cell growth. Nature 249, 123–127.
- 2000. Laser-induced gene expression in specific cells of transgenic zebrafish. Development 127, 1953–1960. , , et al.
- 2000. Proviral insertions in the zebrafish hagoromo gene, encoding an F-box/WD40-repeat protein, cause stripe pattern anomalies. Curr. Biol. 10, 463–466. , , et al.
- 2003. gbx2 Homeobox gene is required for the maintenance of the isthmic region in the zebrafish embryonic brain. Dev. Dyn. 228, 433–450. , , &
- 1995. Stages of embryonic development of the zebrafish. Dev. Dyn. 203, 253–310. , , , &
- 2002. Early subdivisions in the neural plate define distinct competence for inductive signals. Development 129, 83–93. , , , , &
- 1998. Overexpression of the forebrain-specific homeobox gene six3 induces rostral forebrain enlargement in zebrafish. Development 125, 2973–2982. , , , &
- 1991. Expression of the zebrafish paired box gene pax[zf-b] during early neurogenesis. Development 113, 1193–1206. , , &
- 1997. Evidence that FGF8 signalling from the midbrain-hindbrain junction regulates growth and polarity in the developing midbrain. Development 124, 959–969. , , &
- 1998. Generation of destabilized green fluorescent protein as a transcription reporter. J. Biol. Chem. 273, 34 970–34 975. , , et al.
- 1999. FGF8 can activate Gbx2 and transform regions of the rostral mouse brain into a hindbrain fate. Development 126, 4827–4838. , &
- 1995a. FGF-8 isoforms differ in NIH3T3 cell transforming potential. Cell Growth Differ. 6, 817–825. , , , &
- 1995b. FGF-8 isoforms activate receptor splice forms that are expressed in mesenchymal regions of mouse development. Development 121, 3603–3613. , , et al.
- 2000. Regulation and function of FGF8 in patterning of midbrain and anterior hindbrain. Biochem. Cell Biol. 78, 577–584. , , , &
- 2004. Nodal and Fgf pathways interact through a positive regulatory loop and synergize to maintain mesodermal cell populations. Development 131, 629–641. , , et al.
- 1998. An Fgf8 mutant allelic series generated by Cre- and Flp-mediated recombination. Nature Genet 18, 136–141. , &
- 1997. Activator effect of coinjected enhancers on the muscle-specific expression of promoters in zebrafish embryos. Mol. Reprod. Dev. 47, 404–412. , , et al.
- 2002. Modular feedback. Nature 417, 35–36. &
- NIG DNA Sequencing Center 2004. Medaka genome project [database on the Internet]. Cited 27 August 2004. Available from http://dolphin.lab.nig.ac.jp/medaka
- 1998. Zebrafish contains two Pax6 genes involved in eye development. Mec. Dev. 77, 185–196. , , et al.
- 2003. A requirement for retinoic acid-mediated transcriptional activation in ventral neural patterning and motor neuron specification. Neuron 40, 81–95. , , &
- 2006. Structural basis by which alternative splicing modulates the organizer activity of FGF8 in the brain. Genes Dev. 20, 185–198. , ,
- 2001. Fibroblast growth factors. Genome Biol. 2, Reviews 3005. &
- 1993. Cloning of the zebrafish krox-20 gene (krx-20) and its expression during hindbrain development. Nucl Acids Res. 21, 1087–1095. &
- 1998. Fgf8 is mutated in zebrafish acerebellar (ace) mutants and is required for maintenance of midbrain–hindbrain boundary development and somitogenesis. Development 125, 2381–2395. , , , , &
- 2000. Induction and differentiation of the zebrafish heart requires fibroblast growth factor 8 (fgf8/acerebellar). Development 127, 225–235. , , , &
- 2004. Fgf signaling induces posterior neuroectoderm independently of Bmp signaling inhibition. Dev. Dyn. 231, 750–757. , , &
- 2001. The midbrain–hindbrain boundary organizer. Curr. Opin. Neurobiol. 11, 34–42. &
- 2004. Arrays of ultraconserved non-coding regions span the loci of key developmental genes in vertebrate genomes. BMC Genomics. 21, 99. , , et al.
- 2001. Inductive signal and tissue responsiveness defining the tectum and the cerebellum. Development 128, 2461–2469. , &
- 2001. Fgf/MAPK signalling is a crucial positional cue in somite boundary formation. Development 128, 4873–4880. , , , , &
- 1992. The protein product of the zebrafish homologue of the mouse T gene is expressed in nuclei of the germ ring and the notochord of the early embryo. Development 116, 1021–1032. , , &
- 2000. PipMaker – A web server for aligning two genomic DNA sequences Genome Res. 10, 577–586 , , et al.
- 2005. Isolation of Xenopus FGF-8b and comparison with FGF-8a. Mol. Cells 19, 310–317. , , , &
- 1997. Inductive interactions direct early reorganization of the mouse forebrain. Development 124, 2709–2718. &
- 2001. Fgf signalling through MAPK cascade is required for development of the subpallial telencephalon in zebrafish embryos. Development 128, 4153–4164. , , , &
- 2005. Evolutionarily conserved elements in vertebrate, insect, worm, and yeast genomes. Genome Res. 15, 1034–1050. , , et al.
- 2002. Functions of FGF signalling from the apical ectodermal ridge in limb development. Nature 418, 501–508. , &
- 1999. Targeted disruption of Fgf8 causes failure of cell migration in the gastrulating mouse embryo. Genes Dev. 13, 1834–1846. , , &
- 1995. Human androgen-induced growth factor in prostate and breast cancer cells: its molecular cloning and growth properties. FEBS Lett. 363, 226–230. , , , &
- 1992. Cloning and characterization of an androgen-induced growth factor essential for the androgen-dependent growth of mouse mammary carcinoma cells. Proc. Natl Acad. Sci. USA. 89, 8928–8932. , , et al.
- 2004. Hedgehog and Fgf signaling pathways regulate the development of tphR-expressing serotonergic raphe neurons in zebrafish embryos. J. Neurobiol. 60, 275–288. , , et al.
- 2003. Windows into development: historic, current, and future perspectives on transgenic zebrafish. Dev. Biol. 256, 1–17. &
- 2002. Establishment of hindbrain segmental identity requires signaling by FGF3 and FGF8. Curr. Biol. 12, 1117–1123. , , , &
- 2005. Highly conserved non-coding sequences are associated with vertebrate development. PLoS Biol. 3, e7. , , et al.
- 1998. FGF and Shh signals control dopaminergic and serotonergic cell fate in the anterior neural plate. Cell 93, 755–766. , , , &
- 1999. FGF7 and FGF10 directly induce the apical ectodermal ridge in chick embryos. Dev. Biol. 211, 133–143. , , , , &